Device supercell heterojunction

[ams_nanolab]

If we have a heterojunction device supercell (i.e. two different materials A aand B junction at the middle of the central "Device region") as in http://quantumwise.com/support/tutorials/item/94-carbon-nanotube-junctions

and make  a device from Bulk config/ molecular config, what determines whether the electrodes will be considered as periodic? (i.e. is it necessary to have the length of the electrode region to be physically much longer than the central region or vice versa to assume ideal contacts)

Before making the device (i.e. the bulk config) if we repeat the structure in C it becomes A-B-A-B-A-B, what ensures that when we make a device configuration, that it is actually a A-B heterojunction we are simulating.

I just want to know how ATK handles this internally and makes the choice?

[Daniele Stradi]

The electrodes are constructed by internal routines that find periodic patterns starting from the left/right hand side of the central region. A structure like the one you mentioned A-B-A-B-A-B will be transformed in a device (A)-A-B-A-B-A-B-(B), where the letters in parenthesis indicate semi-infinite electrodes, if a periodic pattern can be recognized in A and B.

However, please let me notice that the device configuration is best suited for geometries like L-C-R, where L and R are left and right semi-infinite electrodes and C is a central region where scattering occurs. Check:
http://docs.quantumwise.com/tutorials/atk_transport_calculations/atk_transport_calculations.html

In order to simulate transport in a fully periodic A-B system, I would rather calculate the transmission of the A-B BulkConfiguration.

Regards,
Daniele.

[ams_nanolab]

No in fact I want a (A)-(B) system i.e. both sides semi-infinite. So just making A-B in bulk and then make device from bulk should do the trick ?

I mean (A)-A-B-(B) ?

[Daniele Stradi]

Yes, you are correct.

Just make sure of one thing: a dipole will form at the interface between A and B because of the chemical/physical interaction between the two materials. You have to make sure that this dipole is well screened at the boundary with the electrodes. To do this, just unsure that the A and B parts in the central region are long enough. You can check that the potential is well screened by looking at the Hartree difference potential of the device, which has to be flat close to the electrodes. There is a more technical explanation of this in the tutorial I posted above.

Regards,
Daniele.

[ams_nanolab]

Okay thank you.